
# Basic Libraries 
import numpy as np
import pandas as pd

# Libraries for Visualization
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline 

# Module to Suppress Warnings
from warnings import filterwarnings
filterwarnings('ignore')

# Scikit-learn Modules
from sklearn.model_selection import train_test_split, KFold, cross_val_score, GridSearchCV
from sklearn.linear_model import LinearRegression, LassoCV, Ridge, ARDRegression, BayesianRidge, ElasticNet, ElasticNetCV
from sklearn.linear_model import HuberRegressor, Lars, LarsCV, LassoLarsCV, LassoLarsIC
from sklearn.linear_model import OrthogonalMatchingPursuit, OrthogonalMatchingPursuitCV, PassiveAggressiveRegressor
from sklearn.linear_model import RANSACRegressor, RidgeCV, TweedieRegressor
from sklearn.ensemble import GradientBoostingRegressor, RandomForestRegressor, AdaBoostRegressor, BaggingRegressor
from sklearn.ensemble import ExtraTreesRegressor
from sklearn.tree import DecisionTreeRegressor, ExtraTreeRegressor
from sklearn.cross_decomposition import CCA, PLSRegression
from sklearn.neighbors import KNeighborsRegressor
from sklearn.kernel_ridge import KernelRidge
from sklearn.compose import TransformedTargetRegressor
from sklearn.metrics import explained_variance_score, mean_absolute_error, mean_squared_error, median_absolute_error, r2_score

# Other Modules
from scipy.stats import zscore
from xgboost import XGBRegressor
import pickle


# Pre-setting Plot Style
font={'size':15}
plt.rc('font', **font)
plt.rc('xtick',labelsize=12)    
plt.rc('ytick',labelsize=12)
sns.set_style({'xtick.bottom':True,'ytick.left':True,'text.color':'#9400D3',
               'axes.labelcolor': 'blue','patch.edgecolor': 'black'})

# Settings to display all columns
pd.set_option("display.max_columns", None)


# Loading the data file and creating a dataframe
House_Data = pd.read_csv('innercity.csv')

# Getting Shape and Size of data
H_shape = House_Data.shape

# Displaying the Dataset
print('\033[1mDataset consist:-\033[0m\n Rows    =',H_shape[0],'\n Columns =',H_shape[1],'\n\n\033[1mDataset:-\033[0m')
display(House_Data.head())

Dataset consist:-
 Rows    = 21613 
 Columns = 23 

Dataset:-


# Converting the 'dayhours' attribute to date-time format and keeping Year and Month only
House_Data['dayhours'] = pd.to_datetime(House_Data['dayhours'])

# Checking 'dayhours' attribute after the conversion
print('\n\033[1m"dayhours" attribute after the conversion:-')
display(House_Data[['cid','dayhours']].head())

"dayhours" attribute after the conversion:-


# Exploring dataset
print('\n\033[1mData Types of Each Attribute:-')
display(House_Data.dtypes.to_frame('Data Type'))

Data Types of Each Attribute:-


# Checking for Null Values
print('\n\033[1mNull Values in the Attributes:-')
display(House_Data.isnull().sum().to_frame('Null Values'))

Null Values in the Attributes:-


# Checking for Duplicate Values
print('\n\033[1mDuplicate Values in the Attributes =',House_Data.duplicated().sum())

Duplicate Values in the Attributes = 0


# Describing the data interms of count, mean, standard deviation, and 5 point summary
print('\n\033[1mStatistical Summary of Dataset:-')
display(House_Data.describe().T)

Statistical Summary of Dataset:-


# Checking skewness of the attributes
print('\033[1mSkewness of all attributes:-')  
display(House_Data.skew().to_frame(name='Skewness'))

Skewness of all attributes:-


# Checking Correlation by plotting Heatmap for all attributes 
print('\033[1mHeatmap showing Correlation of Attributes:-')
plt.figure(figsize=(24,22))
plt.title('Correlation of Attributes\n')
sns.heatmap(House_Data.corr(), annot=True, fmt= '.2f', cmap='magma');
plt.show()

Heatmap showing Correlation of Attributes:-


# Creating Plot function for Quantitative Attributes

def qnt_data(x):   
    
    # Plotting Distribution for Quantitative attribute
    print(f'\033[1mPlot Showing Distribution of Feature "{x}":-')
    plt.figure(figsize=(12,6))
    plt.title(f'Distribution of "{x}"\n')
    sns.distplot(House_Data[x],color='#9400D3');
    print('')
    plt.show()
    print('\n__________________________________________________________________________________________________\n')
    print('')
    
    # Box plot for Quantitative data
    print(f'\033[1mPlot Showing 5 point summary with outliers of Attribute "{x}":-\n')
    plt.figure(figsize=(12,6))
    plt.title(f'Box Plot for "{x}"\n')
    sns.boxplot(House_Data[x],color="#9400D3");
    plt.show()

# Creating Plot function for Qualitative Attributes

def qua_data(x):
    
    # Plotting Frequency Distribution of Qualitative attribute
    colors = ['gold','tomato','yellowgreen','#BA4A00','#ADD8E6','#00FF00','#DE3163',]
    print(f'\033[1mPlot Showing Frequency Distribution of Attribute "{x}":-')
    plt.figure(figsize=(10,8))
    plt.title(f'Frequencies of "{x}" Attribute\n')
    sns.countplot(House_Data[x],palette='bright');
    plt.show()
    print('\n___________________________________________________________________________________')
    print('')
    
    # Plotting Pie Chart to check contribution of Qualitative attribute
    print(f'\033[1m\nPie Chart Showing Contribution of Each Category of "{x}" feature:-\n')
    plt.title(f'Contribution of Each Category of "{x}" Attribute\n\n\n\n\n\n\n\n')
    House_Data[x].value_counts().plot.pie(radius=2.5,shadow=True,autopct='%1.1f%%',colors=colors);
    plt.legend(loc='right',prop={'size': 12}, bbox_to_anchor=(2, 1))
    plt.show()


# Univariate analysis for Price Attribute
qnt_data('price')

Plot Showing Distribution of Feature "price":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "price":-


# Univariate analysis for living_measure Attribute
qnt_data('living_measure')

Plot Showing Distribution of Feature "living_measure":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "living_measure":-


# Univariate analysis for lot_measure Attribute
qnt_data('lot_measure')

Plot Showing Distribution of Feature "lot_measure":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "lot_measure":-


# Univariate analysis for ceil_measure Attribute
qnt_data('ceil_measure')

Plot Showing Distribution of Feature "ceil_measure":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "ceil_measure":-


# Univariate analysis for lat Attribute
qnt_data('lat')

Plot Showing Distribution of Feature "lat":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "lat":-


# Univariate analysis for long Attribute
qnt_data('long')

Plot Showing Distribution of Feature "long":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "long":-


# Univariate analysis for Living_measure15 Attribute
qnt_data('living_measure15')

Plot Showing Distribution of Feature "living_measure15":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "living_measure15":-


# Univariate analysis for Lot_measure15 Attribute
qnt_data('lot_measure15')

Plot Showing Distribution of Feature "lot_measure15":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "lot_measure15":-


# Univariate analysis for total_area Attribute
qnt_data('total_area')

Plot Showing Distribution of Feature "total_area":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "total_area":-


# Univariate analysis for yr_built Attribute
qnt_data('yr_built')

Plot Showing Distribution of Feature "yr_built":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "yr_built":-


# Univariate analysis for yr_renovated Attribute
qnt_data('yr_renovated')

Plot Showing Distribution of Feature "yr_renovated":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "yr_renovated":-


# Univariate analysis for basement Attribute
qnt_data('basement')

Plot Showing Distribution of Feature "basement":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "basement":-


# Univariate analysis for coast Attribute
qua_data('coast')

Plot Showing Frequency Distribution of Attribute "coast":-

___________________________________________________________________________________


Pie Chart Showing Contribution of Each Category of "coast" feature:-


# Univariate analysis for Sight Attribute
qua_data('sight')

Plot Showing Frequency Distribution of Attribute "sight":-

___________________________________________________________________________________


Pie Chart Showing Contribution of Each Category of "sight" feature:-


# Univariate analysis for furnished Attribute
qua_data('furnished')
print('\n\033[1m0 = Un-furnished\n1 = Furnished')

Plot Showing Frequency Distribution of Attribute "furnished":-

___________________________________________________________________________________


Pie Chart Showing Contribution of Each Category of "furnished" feature:-

0 = Un-furnished
1 = Furnished


# Univariate analysis for room_bed Attribute
qua_data('room_bed')

Plot Showing Frequency Distribution of Attribute "room_bed":-

___________________________________________________________________________________


Pie Chart Showing Contribution of Each Category of "room_bed" feature:-


# Univariate analysis for room_bath Attribute
qua_data('room_bath')

Plot Showing Frequency Distribution of Attribute "room_bath":-

___________________________________________________________________________________


Pie Chart Showing Contribution of Each Category of "room_bath" feature:-


# Univariate analysis for ceil Attribute
qua_data('ceil')

Plot Showing Frequency Distribution of Attribute "ceil":-

___________________________________________________________________________________


Pie Chart Showing Contribution of Each Category of "ceil" feature:-


# Univariate analysis for condition Attribute
qua_data('condition')

Plot Showing Frequency Distribution of Attribute "condition":-

___________________________________________________________________________________


Pie Chart Showing Contribution of Each Category of "condition" feature:-


# Univariate analysis for quality Attribute
qua_data('quality')

Plot Showing Frequency Distribution of Attribute "quality":-

___________________________________________________________________________________


Pie Chart Showing Contribution of Each Category of "quality" feature:-


# Creating Plot function for Qualitative VS All Quantitative Attribute

def bi_Anly(x):    
    
    # Bivariate Analysis for Qualitative VS All Quantitative Attributes
    print(f'\033[1m\nPlots Showing Bivariate Analysis of "{x}" VS All Quantitative Attributes:-\n')

    # Setting up Sub-Plots
    fig, axes = plt.subplots(6, 2, figsize=(24, 30))
    fig.suptitle(f'"{x}" VS All Quantitative Attributes')
    plt.subplots_adjust(left=0.1,bottom=0.1, right=0.9, top=0.94, wspace=0.3, hspace=0.4)

    # Plotting Sub-Plots
    sns.violinplot(ax=axes[0, 0], x=x, y='yr_built', data=House_Data, palette='bright');
    sns.violinplot(ax=axes[0, 1], x=x, y='yr_renovated', data=House_Data, palette='bright');
    sns.violinplot(ax=axes[1, 0], x=x, y='basement', data=House_Data, palette='bright');
    sns.violinplot(ax=axes[1, 1], x=x, y='price', data=House_Data, palette='bright');
    sns.violinplot(ax=axes[2, 0], x=x, y='living_measure', data=House_Data, palette='bright');
    sns.violinplot(ax=axes[2, 1], x=x, y='lot_measure', data=House_Data, palette='bright');
    sns.violinplot(ax=axes[3, 0], x=x, y='ceil_measure', data=House_Data, palette='bright');
    sns.violinplot(ax=axes[3, 1], x=x, y='lat', data=House_Data, palette='bright');
    sns.violinplot(ax=axes[4, 0], x=x, y='long', data=House_Data, palette='bright');
    sns.violinplot(ax=axes[4, 1], x=x, y='living_measure15', data=House_Data, palette='bright');
    sns.violinplot(ax=axes[5, 0], x=x, y='lot_measure15', data=House_Data, palette='bright');
    sns.violinplot(ax=axes[5, 1], x=x, y='total_area', data=House_Data, palette='bright');

    plt.show()


# Bivariate Analysis for 'room_bed' VS all Quantitative attributes
bi_Anly('room_bed')


Plots Showing Bivariate Analysis of "room_bed" VS All Quantitative Attributes:-


# Bivariate Analysis for 'ceil' VS all Quantitative attributes
bi_Anly('ceil')


Plots Showing Bivariate Analysis of "ceil" VS All Quantitative Attributes:-


# Bivariate Analysis for 'condition' VS all Quantitative attributes
bi_Anly('condition')


Plots Showing Bivariate Analysis of "condition" VS All Quantitative Attributes:-


# Bivariate Analysis for 'quality' VS all Quantitative attributes
bi_Anly('quality')


Plots Showing Bivariate Analysis of "quality" VS All Quantitative Attributes:-


# Bivariate Analysis for 'coast' VS all Quantitative attributes
bi_Anly('coast')


Plots Showing Bivariate Analysis of "coast" VS All Quantitative Attributes:-


# Bivariate Analysis for 'sight' VS all Quantitative attributes
bi_Anly('sight')


Plots Showing Bivariate Analysis of "sight" VS All Quantitative Attributes:-


# Bivariate Analysis for 'furnished' VS all Quantitative attributes
bi_Anly('furnished')


Plots Showing Bivariate Analysis of "furnished" VS All Quantitative Attributes:-


# Plotting for the Multivariate Analysis by considering the target attribute only
print(f'\033[1m\nPlots Showing Multivariate Analysis with respect to target attribute only:-')

# Setting up Sub-Plots
fig, axes = plt.subplots(7, 3, figsize=(30, 50))
fig.suptitle('Price VS All Attributes')
plt.subplots_adjust(left=0.1,bottom=0.1, right=0.97, top=0.96, wspace=0.3, hspace=0.4)

# Plotting Sub-Plots
sns.scatterplot(ax=axes[0, 0], x='room_bed', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[0, 1], x='room_bath', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[0, 2], x='living_measure', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[1, 0], x='lot_measure', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[1, 1], x='ceil', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[1, 2], x='coast', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[2, 0], x='sight', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[2, 1], x='condition', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[2, 2], x='quality', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[3, 0], x='ceil_measure', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[3, 1], x='basement', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[3, 2], x='yr_built', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[4, 0], x='yr_renovated', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[4, 1], x='zipcode', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[4, 2], x='lat', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[5, 0], x='long', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[5, 1], x='living_measure15', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[5, 2], x='lot_measure15', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[6, 0], x='furnished', y='price', data=House_Data, color="#9400D3");
sns.scatterplot(ax=axes[6, 1], x='total_area', y='price', data=House_Data, color="#9400D3");
plt.show()


Plots Showing Multivariate Analysis with respect to target attribute only:-


# Getting the number of houses based on three categories(low,mid and high)
low_PR = len(House_Data[House_Data['price']<500000])

mid_PR = len(House_Data[(House_Data['price']>=500000) & (House_Data['price']<1500000)])

high_PR = len(House_Data[House_Data['price']>=1500000])

# Displaying the results
print('\n\033[1mOn dividing the houses based on Price Range we get:-')
display(pd.DataFrame({'Low':low_PR,'Mid':mid_PR,'High':high_PR},index=['Number of Houses']))

On dividing the houses based on Price Range we get:-


# Getting top 10 Popular Areas
print('\n\033[1mTable showing Top 10 Popular Areas:-')
ZIPcode = House_Data['zipcode'].value_counts().to_frame('Number of Houses')[:10].reset_index()
ZIPcode['ZIPcode'] = ZIPcode['index']
ZIPcode = ZIPcode.drop('index',axis=1).set_index('ZIPcode')

# Displaying the result
display(ZIPcode)

Table showing Top 10 Popular Areas:-


# Getting the Bedroom Trends
print('\n\033[1mTable showing the Bedroom Trends:-')
Bed_Tre = House_Data['room_bed'].value_counts().to_frame('Number of Houses').reset_index()
Bed_Tre['Number of Bedrooms'] = Bed_Tre['index']
Bed_Tre = Bed_Tre.drop('index',axis=1).set_index('Number of Bedrooms')

# Displaying the result
display(Bed_Tre)

Table showing the Bedroom Trends:-


# Getting the Premium Houses based on above given condition
Prem_Hous = House_Data[(House_Data['furnished']==1) & (House_Data['condition']>=3) & (House_Data['total_area']>=10000)]
Prem_Hous = Prem_Hous.sort_values(by=['price'], ascending = False)
# Displaying total number of Premium Houses
print('\n\033[1mTotal Number of Premium Houses =',len(Prem_Hous))
print('\n\033[1mPremium Houses:-')
display(Prem_Hous[['furnished','total_area','condition','room_bed','zipcode','price']].head())

Total Number of Premium Houses = 3013

Premium Houses:-


# Getting Cheapest Houses sold in Coastal areas
Costl_Hous = House_Data[(House_Data['coast']==1) & (House_Data['price']<=500000)]

# Displaying total number of Cheapest Houses in Coastal areas
print('\n\033[1mTotal Number of Cheapest Houses in Coastal areas =',len(Costl_Hous))
print('\n\033[1mCheapest Coastal Houses:-')
display(Costl_Hous[['cid','coast','price']])

Total Number of Cheapest Houses in Coastal areas = 14

Cheapest Coastal Houses:-


Non_Reno = House_Data[(House_Data['yr_renovated']==0)]['price'].value_counts()
Reno = House_Data[(House_Data['yr_renovated']!=0)]['price'].value_counts()

# Displaying  Price pattern of Renovated and Non-renovated houses
print('\n\033[1mTotal number of Renovated Houses =',len(House_Data[(House_Data['yr_renovated']!=0)]))
print('\n\033[1mTotal number of Non-Renovated Houses =',len(House_Data[(House_Data['yr_renovated']==0)]))
NrnRn = {'Renovated Price':sorted(Reno.index[:10],reverse=True),'Non Renovated Price':sorted(Non_Reno.index[:10],reverse=True)}
print('\n\033[1mTable showing top 10 Price Comaprison between Renovated and Non-renovated houses:-')
display(pd.DataFrame(NrnRn))

Total number of Renovated Houses = 914

Total number of Non-Renovated Houses = 20699

Table showing top 10 Price Comaprison between Renovated and Non-renovated houses:-


# Getting the Month in which buyers bought more number of houses
Month_hous = House_Data['dayhours'].dt.month.value_counts().to_frame('Number of Houses').reset_index()
Month_hous['Month'] = Month_hous['index']
Month_hous = Month_hous.drop('index',axis=1).set_index('Month')

# Displaying the result
print('\n\033[1mThe Month in which buyers bought more number of houses is =',int(Month_hous.index[:1].values))
print('\n\033[1mTable showing the Months in which buyers bought houses.:-')
display(Month_hous)

The Month in which buyers bought more number of houses is = 5

Table showing the Months in which buyers bought houses.:-


# Displaying relation between Year_built and Price
print(f'\033[1mPlot Showing relation between Year_built and Price :-')
plt.figure(figsize=(12,6))
plt.title('Year_built VS Price \n')
sns.lineplot(x='yr_built', y='price', data=House_Data, color="#9400D3");
plt.show()

Plot Showing relation between Year_built and Price :-


# Dropping customerID Attribute
House_Data.drop(['cid','dayhours'],axis=1,inplace=True)

# Getting Shape and Size of data
H_shape = House_Data.shape

# Checking the Dataset
print('\033[1mDataset consist:-\033[0m\n Rows    =',H_shape[0],'\n Columns =',H_shape[1],'\n\n\033[1mDataset:-\033[0m')
display(House_Data.head())

Dataset consist:-
 Rows    = 21613 
 Columns = 21 

Dataset:-


# Creating list of Continuous Features
continuous_features = ['price','living_measure','lot_measure','ceil_measure','lat','long','living_measure15','lot_measure15',
                     'total_area']

# Creating empty lists
Attributes = []   # to store attribute names
OL_before = []    # to store outliers before flooring and capping
OL_after = []     # to store outliers after flooring and capping

# Sending all the Continuous columns inside the loop
for i in continuous_features:
    Attributes.append(i)
    
    # Getting Interquartile Range
    q1 = House_Data[i].quantile(0.25)
    q3 = House_Data[i].quantile(0.75)
    IQR = q3 - q1
    
    # Getting Outliers Before Flooring and Capping Imputation
    cnt1 = 0   # counter to count the outliers
    for k in House_Data[i]:
        if (k < (q1 - 1.5 * IQR) or k > (q3 + 1.5 * IQR)):
            cnt1 += 1
    OL_before.append(cnt1)
    
    # Getting Lower fence and Upper fence
    LF = q1 - (1.5*IQR)
    UF = q3 + (1.5*IQR)
    
    # Replacing all the outliers by upper and lower fence values
    House_Data[i] = np.where(House_Data[i]>UF,UF,np.where(House_Data[i]<LF,LF,House_Data[i]))
    
    # Getting Outliers After Flooring and Capping Imputation
    cnt2 = 0    # counter to count the outliers
    for k in House_Data[i]:
        if (k < (q1 - 1.5 * IQR) or k > (q3 + 1.5 * IQR)):
            cnt2 += 1
    OL_after.append(cnt2)
    
# Creting dataframe for better representation of Outlier Analysis      
Outlier_Analysis = pd.DataFrame({'Attribute':Attributes,
                                 'Outliers Before F&C':OL_before,
                                 'Outliers After F&C':OL_after})

# Displaying Outlier Analysis report
print('\n\033[1mOutlier Analysis Report:-')
display(Outlier_Analysis) 
print('\033[1mTotal Outliers Observed in the all the attributes together =',sum(OL_before))

# Checking the distribution for Price Attribute
print('\n\nChecking the distribution for Price Attribute:')
qnt_data('price')

Outlier Analysis Report:-

Total Outliers Observed in the all the attributes together = 10182


Checking the distribution for Price Attribute:
Plot Showing Distribution of Feature "price":-

__________________________________________________________________________________________________


Plot Showing 5 point summary with outliers of Attribute "price":-


# Creating list of Continuous Features
continuous_features = ['price','living_measure','lot_measure','ceil_measure','lat','long','living_measure15','lot_measure15',
                     'total_area']

# Storing price value(without scaling) for future use
Input_Data = House_Data.drop(columns='price')
Price = House_Data['price'].copy()

# Applying Z-Scores to the continuous data attributes
House_Data[continuous_features] = House_Data[continuous_features].apply(zscore)

# Checking the Mean and Standard Deviation
print('\033[1mTable Showing Mean and Standard Deviation of Scaled Continuous Data:-')
display(House_Data[continuous_features].describe()[1:3].T)

Table Showing Mean and Standard Deviation of Scaled Continuous Data:-


# Getting Predictors by dropping Class Attribute
X = House_Data.drop(columns='price')

# Getting Target Attribute
y = House_Data['price'] 

print('\033[1mTable Showing Segregated Predictors:-')
display(X.head())

Table Showing Segregated Predictors:-


# Splitting into Train and Test Sets with the ratio 75:25
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=1, test_size=0.25)

# For better observation of Splitted Data
TTS = pd.DataFrame({'Data':['X_train','y_train','X_test','y_test'],
                    'Rows':[len(X_train),len(y_train),len(X_test),len(y_test)],
                    'Columns':[len(X_train.columns),1,len(X_test.columns),1]}).set_index('Data')
print('\033[1mTable Showing splitted data into Train and Test sets:-')
display(TTS)

Table Showing splitted data into Train and Test sets:-


# Creating list of Models for our data
Models = [LinearRegression(), LassoCV(), Ridge(), ARDRegression(), BayesianRidge(), ElasticNet(), ElasticNetCV(),  RidgeCV(),
          HuberRegressor(), Lars(), LarsCV(), LassoLarsCV(), LassoLarsIC(), OrthogonalMatchingPursuit(), XGBRegressor(),
          OrthogonalMatchingPursuitCV(), RANSACRegressor(), TweedieRegressor(),GradientBoostingRegressor(),
          RandomForestRegressor(), AdaBoostRegressor(), BaggingRegressor(), ExtraTreesRegressor(),DecisionTreeRegressor(), 
          ExtraTreeRegressor(), CCA(), PLSRegression(), KNeighborsRegressor(), KernelRidge(), TransformedTargetRegressor()]
 
# Creating empty sets to store train and test accracies
Train_Accuracy = []
Test_Accuracy = []
Diff_Accuracy = []

# Training models and getting accuracies for Train and Test Data
for model in Models:
    model.fit(X_train,y_train)
    train = round(model.score(X_train,y_train),4)*100
    test = round(model.score(X_test,y_test),4)*100
    Train_Accuracy.append(train)
    Test_Accuracy.append(test)
    Diff_Accuracy.append(train-test)

# Getting differences between train and test accuracies

# Getting model names only for better representation
model_names = [x[:x.index('(')] for x in pd.Series(Models).apply(str)]     
    
# Displaying the Model Accuracy Report
print('\033[1mAccuracy Report of all the Models:-')
Model_Report1 = pd.DataFrame({'Models':model_names, 'Train Accuracy (%)':Train_Accuracy, 'Test Accuracy (%)':Test_Accuracy,
                              '(Train Accuracy - Test Accuracy)':Diff_Accuracy})
Model_Report1 = Model_Report1.sort_values(['Test Accuracy (%)'],ascending=False).reset_index().drop('index',axis=1)
display(Model_Report1)

Accuracy Report of all the Models:-


# Getting Feature Importances from varius techniques

# Extra Tree Regression Feature Importance 
ET = ExtraTreesRegressor()
ET.fit(X, y)             
ET_imp = ET.feature_importances_  # Getting feature importance
ET = pd.Series(ET_imp, index=X.columns).sort_values()

# Random Forest Feature Importance
RF = RandomForestRegressor(random_state=1)
RF.fit(X, y)
RF_imp = RF.feature_importances_
RF = pd.Series(RF_imp, index=X.columns).sort_values()

# XGBoost Feature Importance
XG = XGBRegressor(random_state=1)
XG.fit(X, y)
XG_imp = XG.feature_importances_
XG = pd.Series(XG_imp, index=X.columns).sort_values()

# Bagging Regressor Feature Importance
BG = BaggingRegressor(random_state=1)
BG.fit(X, y)
BG_imp = np.mean([
    tree.feature_importances_ for tree in BG.estimators_
], axis=0)
BG = pd.Series(BG_imp, index=X.columns).sort_values()

# GradientBoost Feature Importance
GB =GradientBoostingRegressor(random_state=1)
GB.fit(X, y)
GB_imp = GB.feature_importances_
GB = pd.Series(GB_imp, index=X.columns).sort_values()

# Decision Tree Feature Importance
DT = DecisionTreeRegressor(random_state=1)
DT.fit(X, y)
DT_imp = DT.feature_importances_
DT = pd.Series(DT_imp, index=X.columns).sort_values()

# Displaying the  Feature Importances shown by varius techniques

# Getting subplots

fig, axis = plt.subplots(3, 2, figsize=(24, 30)) 
plt.subplots_adjust(wspace=0.25, hspace=0.3, left=0.1)

# Plotting Feature Importances
ET.plot(kind='barh', color= '#9400D3',ax=axis[0, 0])  
axis[0, 0].set_title('Extra Tree Regression Feature Importance\n')

RF.plot(kind='barh', color= '#9400D3',ax=axis[0, 1])
axis[0, 1].set_title('Random Forest Feature Importance\n')

XG.plot(kind='barh', color= '#9400D3',ax=axis[1, 0])
axis[1, 0].set_title('XGBoost Feature Importance\n')

BG.plot(kind='barh', color= '#9400D3',ax=axis[1, 1])
axis[1, 1].set_title('Bagging Feature Importance\n')

GB.plot(kind='barh', color= '#9400D3',ax=axis[2, 0])
axis[2, 0].set_title('GradientBoost Feature Importance\n')


DT.plot(kind='barh', color= '#9400D3',ax=axis[2, 1])
axis[2, 1].set_title('Decision Tree Feature Importance\n')

print('\033[1mPlots showing Feature Importances shown by varius techniques:-\n')
plt.show()

Plots showing Feature Importances shown by varius techniques:-


# Creating a list of feature importance techniques and empty list to store accuracies and number of columns
Fea_imp = [ET,DT,RF,XG,BG,GB]
Train_Accuracy = []
Test_Accuracy = []
Columns = []

# Getting the accuracies based on different threshold values
for top_scr in range(0,20):
    
    # Segregating the data
    x = House_Data.drop(columns='price')
    Y = House_Data['price']
    
    # Creating empty list to store top features based on various threshold value
    Top_Features = []
    
    for col in x.columns:
        for imp_scr in Fea_imp:
            if col in imp_scr[top_scr:] and col not in Top_Features:
                Top_Features.append(col) # Appending top features according to the condition
     
    # Considering only the top features in predictors
    x = x[Top_Features]
    
    # Getting number of columns of predictors
    Columns.append(x.shape[1])
    
    # Splitting the data to train and test sets in 75:25 ratio
    X_train, X_test, y_train, y_test = train_test_split(x, Y, random_state=1, test_size=0.25)
    
    # Building the model
    mdl = GradientBoostingRegressor(random_state=1)
    mdl.fit(X_train,y_train)
    
    # Getting the accuracies
    Train_Accuracy.append(round(mdl.score(X_train,y_train),4)*100)
    Test_Accuracy.append(round(mdl.score(X_test,y_test),4)*100)
    
# Displaying Threshold Report
Threshold_Report = pd.DataFrame({'Threshold':list(range(0,20)),'Number of Columns':Columns,'Train Accuracy':Train_Accuracy,
                                 'Test Accuracy':Test_Accuracy}).set_index('Threshold')
print('\033[1mTable Showing Accuracies based on Threshold values:-')
display(Threshold_Report.sort_values(['Test Accuracy'],ascending=False))

Table Showing Accuracies based on Threshold values:-


# Creating a list of feature importance techniques and empty list to store selected features
Fea_imp = [ET,DT,RF,XG,BG,GB]
Top_Features = []
threshold = 8  # Setting the threshold as 8

# Getting the features based threshold score
for col in X.columns:
    for imp_scr in Fea_imp:
        if col in imp_scr[threshold:] and col not in Top_Features: # Selecting features that appear on top 8 scores
            Top_Features.append(col)

# Selecting the attributes(feature selection) based on threshold value
X = X[Top_Features]

# Splitting the data to train and test sets with 75:25 ratio
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=1, test_size=0.25)

# Getting Shape and Size of Predictor data 
X_shape = X.shape

# Checking the Dataset
print('\033[1mPredictor data consist:-\033[0m\n Rows    =',X_shape[0],'\n Columns =',X_shape[1],'\n\n\033[1mDataset:-\033[0m')
display(X.head())

Predictor data consist:-
 Rows    = 21613 
 Columns = 16 

Dataset:-


# Getting accuracies without tuning and before reducing features
model_1 = GradientBoostingRegressor(random_state=1)
model_1.fit(X_train,y_train)
train_1 = round(model_1.score(X_train,y_train),4)*100
test_1 = round(model_1.score(X_test,y_test),4)*100

# Getting accuracies with tuning and before reducing features
model_2 = GradientBoostingRegressor(random_state=1, learning_rate=0.3, subsample=0.7, min_samples_leaf=40)
model_2.fit(X_train,y_train)
train_2 = round(model_2.score(X_train,y_train),4)*100
test_2 = round(model_2.score(X_test,y_test),4)*100

# Reducing Features and splitting the data to train and test sets
X = X.drop(['coast','furnished','lot_measure'],axis=1) 

# After checking changes in accuracies we dropped these features
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=1, test_size=0.25) 

# Getting accuracies without tuning and after reducing features
model_3 = GradientBoostingRegressor(random_state=1)
model_3.fit(X_train,y_train)
train_3 = round(model_3.score(X_train,y_train),4)*100
test_3 = round(model_3.score(X_test,y_test),4)*100

# Getting accuracies with tuning and after reducing features
model_4 = GradientBoostingRegressor(random_state=1, learning_rate=0.3, subsample=0.7, min_samples_leaf=40)
model_4.fit(X_train,y_train)
train_4 = round(model_4.score(X_train,y_train),4)*100
test_4 = round(model_4.score(X_test,y_test),4)*100

# Displaying Accuracies of GradientBoostingRegressor
print('\033[1mReport showing Accuracy and RMSE of GradientBoostingRegressor model:-')
Model_Report3 = pd.DataFrame({'Feature Reduction':['Before','After'],'Train ':[train_1,train_3],'Test':[test_1, test_3],
                              'Train':[train_2,train_4],'Test ':[test_2,test_4]}).set_index('Feature Reduction')
columns=[('Accuracy Without Tuning (%)', 'Train '),('Accuracy Without Tuning (%)', 'Test'),
         ('Accuracy With Tuning (%)', 'Train'),('Accuracy With Tuning (%)', 'Test ')]
Model_Report3.columns = pd.MultiIndex.from_tuples(columns)
display(Model_Report3)

# Displaying the final shape of predictors
print('\n\n\033[1mPredictors after feature reduction has:-\033[0m\n Rows    =',X.shape[0],'\n Columns =',X.shape[1])

Report showing Accuracy and RMSE of GradientBoostingRegressor model:-


Predictors after feature reduction has:-
 Rows    = 21613 
 Columns = 13


# Fitting the model
Model = GradientBoostingRegressor(random_state=1, learning_rate=0.3, subsample=0.7, min_samples_leaf=40)
Model.fit(X_train,y_train)
y_pred = Model.predict(X_test)

# Getting Train and Test Accuracies
Train = round(Model.score(X_train,y_train),4)*100
Test = round(Model.score(X_test,y_test),4)*100

# Getting variance score, error scores and r2 value
EV = explained_variance_score(y_test, y_pred)
MAE = mean_absolute_error(y_test, y_pred) 
MSE = mean_squared_error(y_test, y_pred) 
MdAE = median_absolute_error(y_test, y_pred)
R2 = r2_score(y_test, y_pred)

# Displaying the result
shape = X.shape
print('\033[1mFinalised Dataset consist:-','\n Rows =',shape[0],'\n Columns =',shape[1],'\n Target Attribute = Price',
      '\n\n\033[1mDataset:-\033[0m')
display(X.head())

parameters = ['Model Name','Train Accuracy (%)','Test Accuracy (%)','R-squared value','Explained Variance',
              'Mean Absolute Error','Mean Squared Error','Median Absolute Error','Root Mean Square Error']
Results = ['GradientBoostingRegressor',Train,Test,round(R2,4),round(EV,4),round(MAE,4),round(MSE,4),round(MdAE,4),
           round(np.sqrt(MSE),4)]
Final_Report = pd.DataFrame({'Parameters':parameters,'Result':Results})
print('\033[1m\n\nReport showing the results of our Final Model:-')
display(Final_Report)

Finalised Dataset consist:- 
 Rows = 21613 
 Columns = 13 
 Target Attribute = Price 

Dataset:-



Report showing the results of our Final Model:-


# Saving the Final best trained model to the file
pickle.dump(Model, open('Housing_Model.pkl', 'wb')) 

# Loading the Pickled File and getting the model
pickled_file = open('Housing_Model.pkl', 'rb')
pickled_model = pickle.load(pickled_file) 
pickled_file.close()
  
# Checking the loaded model by finding accuracies
Train_AC = round(pickled_model.score(X_train , y_train),4)*100
Test_AC = round(pickled_model.score(X_test , y_test),4)*100

print('\033[1mTable Showing Pickled Model Accuracies for Train and Test Data:-')
display(pd.DataFrame({'Data':['Training','Testing'],'Model Accuracy (%)':
                      [Train_AC,Test_AC]}).set_index('Data'))

Table Showing Pickled Model Accuracies for Train and Test Data:-


# Creating New Data
raw_data = Input_Data[X.columns][0:5]  # For displaying raw input before scaling it
New_Data = X[0:5]   # Scaled version of raw data (pre-scaled)

# Generating Prediction
Z_pred = pickled_model.predict(New_Data)  # Returns z-scored price in array

# Re-transforming predicted values from z-scored to normal
sd = Price.std()
m = Price.mean()
Prediction = Z_pred*sd + m  # Re-transforming using Z-score formula (z = (x-μ)/σ)

# Displaying the Prediction Report
print('\033[1mPrediction Report:-\n')
print('\033[1m Input Data:')
display(raw_data)
print('\033[1m\n Prediction Report:')
display(pd.DataFrame({'Actual Price':Price[0:5],'Predicted Price':np.round(Prediction),
                      'Difference':Price[0:5] - np.round(Prediction),
                      'HL Percentage(%)':round(((Price[0:5] - np.round(Prediction))/Price[0:5]*100),2),
                      'Prediction Accuracy (%)':round(np.round(Prediction)/Price[0:5]*100,2)}))

Prediction Report:-

 Input Data:


 Prediction Report:

	cid	dayhours	price	room_bed	room_bath	living_measure	lot_measure	ceil	condition	quality	ceil_measure	basement	yr_built	zipcode	lat	long	living_measure15	lot_measure15	furnished	total_area
0	3034200666	20141107T000000	808100	4	3.25	3020	13457	1.0	5	9	3020	0	1956	98133	47.7174	-122.336	2120	7553	1	16477
1	8731981640	20141204T000000	277500	4	2.50	2550	7500	1.0	3	8	1750	800	1976	98023	47.3165	-122.386	2260	8800	0	10050
2	5104530220	20150420T000000	404000	3	2.50	2370	4324	2.0	3	8	2370	0	2006	98038	47.3515	-121.999	2370	4348	0	6694
3	6145600285	20140529T000000	300000	2	1.00	820	3844	1.0	4	6	820	0	1916	98133	47.7049	-122.349	1520	3844	0	4664
4	8924100111	20150424T000000	699000	2	1.50	1400	4050	1.0	4	8	1400	0	1954	98115	47.6768	-122.269	1900	5940	0	5450

	cid	dayhours
0	3034200666	2014-11-07
1	8731981640	2014-12-04
2	5104530220	2015-04-20
3	6145600285	2014-05-29
4	8924100111	2015-04-24

	count	mean	std	min	25%	50%	75%	max
cid	21613.0	4.580302e+09	2.876566e+09	1.000102e+06	2.123049e+09	3.904930e+09	7.308900e+09	9.900000e+09
price	21613.0	5.401822e+05	3.673622e+05	7.500000e+04	3.219500e+05	4.500000e+05	6.450000e+05	7.700000e+06
room_bed	21613.0	3.370842e+00	9.300618e-01	0.000000e+00	3.000000e+00	3.000000e+00	4.000000e+00	3.300000e+01
room_bath	21613.0	2.114757e+00	7.701632e-01	0.000000e+00	1.750000e+00	2.250000e+00	2.500000e+00	8.000000e+00
living_measure	21613.0	2.079900e+03	9.184409e+02	2.900000e+02	1.427000e+03	1.910000e+03	2.550000e+03	1.354000e+04
lot_measure	21613.0	1.510697e+04	4.142051e+04	5.200000e+02	5.040000e+03	7.618000e+03	1.068800e+04	1.651359e+06
ceil	21613.0	1.494309e+00	5.399889e-01	1.000000e+00	1.000000e+00	1.500000e+00	2.000000e+00	3.500000e+00
coast	21613.0	7.541757e-03	8.651720e-02	0.000000e+00	0.000000e+00	0.000000e+00	0.000000e+00	1.000000e+00
sight	21613.0	2.343034e-01	7.663176e-01	0.000000e+00	0.000000e+00	0.000000e+00	0.000000e+00	4.000000e+00
condition	21613.0	3.409430e+00	6.507430e-01	1.000000e+00	3.000000e+00	3.000000e+00	4.000000e+00	5.000000e+00
quality	21613.0	7.656873e+00	1.175459e+00	1.000000e+00	7.000000e+00	7.000000e+00	8.000000e+00	1.300000e+01
ceil_measure	21613.0	1.788391e+03	8.280910e+02	2.900000e+02	1.190000e+03	1.560000e+03	2.210000e+03	9.410000e+03
basement	21613.0	2.915090e+02	4.425750e+02	0.000000e+00	0.000000e+00	0.000000e+00	5.600000e+02	4.820000e+03
yr_built	21613.0	1.971005e+03	2.937341e+01	1.900000e+03	1.951000e+03	1.975000e+03	1.997000e+03	2.015000e+03
yr_renovated	21613.0	8.440226e+01	4.016792e+02	0.000000e+00	0.000000e+00	0.000000e+00	0.000000e+00	2.015000e+03
zipcode	21613.0	9.807794e+04	5.350503e+01	9.800100e+04	9.803300e+04	9.806500e+04	9.811800e+04	9.819900e+04
lat	21613.0	4.756005e+01	1.385637e-01	4.715590e+01	4.747100e+01	4.757180e+01	4.767800e+01	4.777760e+01
long	21613.0	-1.222139e+02	1.408283e-01	-1.225190e+02	-1.223280e+02	-1.222300e+02	-1.221250e+02	-1.213150e+02
living_measure15	21613.0	1.986552e+03	6.853913e+02	3.990000e+02	1.490000e+03	1.840000e+03	2.360000e+03	6.210000e+03
lot_measure15	21613.0	1.276846e+04	2.730418e+04	6.510000e+02	5.100000e+03	7.620000e+03	1.008300e+04	8.712000e+05
furnished	21613.0	1.966872e-01	3.975030e-01	0.000000e+00	0.000000e+00	0.000000e+00	0.000000e+00	1.000000e+00
total_area	21613.0	1.718687e+04	4.158908e+04	1.423000e+03	7.035000e+03	9.575000e+03	1.300000e+04	1.652659e+06

	Number of Houses
ZIPcode
98103	602
98038	590
98115	583
98052	574
98117	553
98042	548
98034	545
98118	508
98023	499
98006	498

	Number of Houses
Number of Bedrooms
3	9824
4	6882
2	2760
5	1601
6	272
1	199
7	38
8	13
0	13
9	6
10	3
11	1
33	1

	Data Type
cid	int64
dayhours	datetime64[ns]
price	int64
room_bed	int64
room_bath	float64
living_measure	int64
lot_measure	int64
ceil	float64
coast	int64
sight	int64
condition	int64
quality	int64
ceil_measure	int64
basement	int64
yr_built	int64
yr_renovated	int64
zipcode	int64
lat	float64
long	float64
living_measure15	int64
lot_measure15	int64
furnished	int64
total_area	int64

	Skewness
cid	0.243329
price	4.021716
room_bed	1.974300
room_bath	0.511108
living_measure	1.471555
lot_measure	13.060019
ceil	0.616177
coast	11.385108
sight	3.395750
condition	1.032805
quality	0.771103
ceil_measure	1.446664
basement	1.577965
yr_built	-0.469805
yr_renovated	4.549493
zipcode	0.405661
lat	-0.485270
long	0.885053
living_measure15	1.108181
lot_measure15	9.506743
furnished	1.526232
total_area	12.956067

	furnished	total_area	condition	room_bed	zipcode	price
16785	1	39650	4	6	98102	7700000
19888	1	47365	3	5	98004	7060000
264	1	41264	3	6	98039	6890000
20917	1	44269	3	5	98039	5570000
17618	1	31985	3	5	98004	5350000

	cid	coast	price
2040	221029019	1	400000
4383	8550001515	1	429592
5993	2013802060	1	500000
7258	2423029009	1	465000
8862	2923039243	1	340000
11194	2123039032	1	369900
12120	3523029041	1	290000
12669	222029026	1	340000
13157	2013802030	1	357000
14514	7631800110	1	380000
14897	121039042	1	425000
15273	5216200090	1	385000
18189	1922039062	1	480000
19672	2781600195	1	285000

	Renovated Price	Non Renovated Price
0	1100000	550000
1	900000	500000
2	850000	450000
3	800000	425000
4	750000	400000
5	700000	375000
6	550000	350000
7	500000	325000
8	450000	300000
9	400000	250000

	Number of Houses
Month
5	2414
4	2231
7	2211
6	2180
8	1940
10	1878
3	1875
9	1774
12	1471
11	1411
2	1250
1	978

	Attribute	Outliers Before F&C
0	price	1159
1	living_measure	572
2	lot_measure	2425
3	ceil_measure	611
4	lat	2
5	long	256
6	living_measure15	544
7	lot_measure15	2194
8	total_area	2419

	mean	std
price	4.667324e-17	1.000023
living_measure	1.743235e-16	1.000023
lot_measure	1.064197e-16	1.000023
ceil_measure	9.062395e-17	1.000023
lat	-2.431033e-14	1.000023
long	3.689728e-14	1.000023
living_measure15	5.395726e-17	1.000023
lot_measure15	1.340918e-16	1.000023
total_area	-6.541239e-17	1.000023

	room_bed	room_bath	living_measure	lot_measure	ceil	condition	quality	ceil_measure	basement	yr_built	zipcode	lat	long	living_measure15	lot_measure15	furnished	total_area
0	4	3.25	1.146116	0.941623	1.0	5	9	1.636673	0	1956	98133	1.135591	-0.882558	0.223252	-0.170774	1	1.050860
1	4	2.50	0.586117	-0.238830	1.0	3	8	-0.025606	800	1976	98023	-1.757746	-1.246974	0.438974	0.115013	0	-0.135220
2	3	2.50	0.371650	-0.868194	2.0	3	8	0.785900	0	2006	98038	-1.505147	1.573604	0.608470	-0.905297	0	-0.754558
3	2	1.00	-1.475153	-0.963312	1.0	4	6	-1.242865	0	1916	98133	1.045377	-0.977306	-0.701270	-1.020804	0	-1.129188
4	2	1.50	-0.784091	-0.922491	1.0	4	8	-0.483714	0	1954	98115	0.842577	-0.394241	-0.115739	-0.540442	0	-0.984134

	Models	Train Accuracy (%)	Test Accuracy (%)	(Train Accuracy - Test Accuracy)
0	ExtraTreesRegressor	99.85	89.63	10.22
1	RandomForestRegressor	98.39	89.49	8.90
2	XGBRegressor	96.25	89.47	6.78
3	BaggingRegressor	97.73	88.40	9.33
4	GradientBoostingRegressor	88.76	87.65	1.11
5	DecisionTreeRegressor	99.85	78.10	21.75
6	ExtraTreeRegressor	99.85	76.23	23.62
7	LinearRegression	76.29	75.83	0.46
8	LassoLarsCV	76.29	75.83	0.46
9	LassoLarsIC	76.29	75.83	0.46
10	TransformedTargetRegressor	76.29	75.83	0.46
11	LarsCV	76.29	75.83	0.46
12	Lars	76.29	75.83	0.46
13	Ridge	76.29	75.83	0.46
14	RidgeCV	76.29	75.83	0.46
15	BayesianRidge	76.29	75.82	0.47
16	KernelRidge	75.91	75.44	0.47
17	CCA	75.75	75.21	0.54
18	ARDRegression	73.38	73.31	0.07
19	OrthogonalMatchingPursuitCV	73.60	73.21	0.39
20	PLSRegression	72.46	71.56	0.90
21	HuberRegressor	71.90	71.16	0.74
22	LassoCV	68.37	67.97	0.40
23	TweedieRegressor	68.75	67.89	0.86
24	AdaBoostRegressor	67.89	66.83	1.06
25	ElasticNetCV	67.21	66.71	0.50
26	RANSACRegressor	65.15	66.21	-1.06
27	OrthogonalMatchingPursuit	60.23	60.70	-0.47
28	KNeighborsRegressor	71.77	56.23	15.54
29	ElasticNet	29.51	29.15	0.36

	Rows	Columns
Data
X_train	16209	20
y_train	16209	1
X_test	5404	20
y_test	5404	1

	Number of Columns	Train Accuracy	Test Accuracy
Threshold
8	16	88.76	87.78
9	14	88.69	87.77
10	13	88.58	87.73
4	18	88.74	87.66
5	18	88.74	87.66
6	18	88.74	87.66
7	18	88.74	87.66
1	20	88.76	87.65
0	20	88.76	87.65
3	19	88.76	87.65
2	20	88.76	87.65
11	12	88.32	87.47
12	12	88.32	87.47
13	10	87.84	86.97
14	9	87.63	86.66
15	8	87.33	86.65
16	6	86.04	85.50
17	4	80.83	80.49
18	4	80.83	80.49
19	3	72.84	72.24

	Parameters	Result
0	Model Name	GradientBoostingRegressor
1	Train Accuracy (%)	90.42
2	Test Accuracy (%)	89.32
3	R-squared value	0.8932
4	Explained Variance	0.8932
5	Mean Absolute Error	0.2332
6	Mean Squared Error	0.1059
7	Median Absolute Error	0.1702
8	Root Mean Square Error	0.3254

	Actual Price	Predicted Price	Difference	HL Percentage(%)	Prediction Accuracy (%)
0	808100.0	797173.0	10927.0	1.35	98.65
1	277500.0	284683.0	-7183.0	-2.59	102.59
2	404000.0	374064.0	29936.0	7.41	92.59
3	300000.0	295664.0	4336.0	1.45	98.55
4	699000.0	622997.0	76003.0	10.87	89.13

	Accuracy Without Tuning (%)		Accuracy With Tuning (%)
	Train	Test	Train	Test
Feature Reduction
Before	88.76	87.78	90.48	89.09
After	88.56	87.71	90.42	89.32

	room_bath	living_measure	condition	quality	ceil_measure	yr_built	zipcode	lat	long	living_measure15	lot_measure15	total_area
0	3.25	3020.0	5	9	3020.0	1956	98133	47.7174	-122.336	2120.0	7553.0	16477.0
1	2.50	2550.0	3	8	1750.0	1976	98023	47.3165	-122.386	2260.0	8800.0	10050.0
2	2.50	2370.0	3	8	2370.0	2006	98038	47.3515	-121.999	2370.0	4348.0	6694.0
3	1.00	820.0	4	6	820.0	1916	98133	47.7049	-122.349	1520.0	3844.0	4664.0
4	1.50	1400.0	4	8	1400.0	1954	98115	47.6768	-122.269	1900.0	5940.0	5450.0

CAPSTONE PROJECT MACHINE LEARNING¶

Importing Required Python Libraries and Modules

Import and Warehouse Data

- Exploring the Data

Data Cleansing

- Modification of the Attributes

- Exploring Datatypes of Each Attribute

- Checking for Null Values in the Attributes

- Checking for Duplicate Values

Data Analysis & Visualisation:

Information about the Features¶

1. Qualitative Attributes:-¶

2. Quantitative Attributes:-¶

- Statistical Summary of Data

- Checking Skewness of the Attributes

- Checking Correlation of the Attributes

- Univariate Analysis

Creating Functions for Plotting the Quantitative and Qualitative Attributes for Univariate Analysis.

Attribute 1: Price¶

Attribute 2: Living_Measure¶

Attribute 3: Lot_Measure¶

Attribute 4: Ceil_Measure¶

Attribute 5: Lat¶

Attribute 6: Long¶

Attribute 7: Living_measure15¶

Attribute 8: Lot_measure15¶

Attribute 9: Total_Area¶

Attribute 10: Yr_Built¶

Attribute 11: Yr_Renovated¶

Attribute 12: Basement¶

Attribute 13: Coast¶

Attribute 14: Sight¶

Attribute 15: Furnished¶

Attribute 16: Room_bed¶

Attribute 17: Room_bath¶

Attribute 18: Ceil¶

Attribute 19: Condition¶

Attribute 20: Quality¶

- Bivariate Analysis

Creating Functions for Plotting the Quantitative VS Categorical Data for Bivariate Analysis

Bivariate Analysis 1: 'room_bed' VS all Quantitative attributes¶

Bivariate Analysis 2: 'ceil' VS all Quantitative attributes¶

Bivariate Analysis 3: 'condition' VS all Quantitative attributes¶

Bivariate Analysis 4: 'quality' VS all Quantitative attributes¶

Bivariate Analysis 5: 'coast' VS all Quantitative attributes¶

Bivariate Analysis 6: 'sight' VS all Quantitative attributes¶

Bivariate Analysis 7: 'furnished' VS all Quantitative attributes¶

- Multivariate Analysis

Finding the Hidden Patterns

- Dividing the houses based on Price Range(PR)

- Finding Top 10 Popular areas where the Buyers shown their interest to Buy Properties

- Finding the Bedroom Trends

- Finding Premium Houses having following requirement:-

- Cheapest Houses sold in Coastal areas

- Checking the Price Comaprison between Renovated and Non-renovated houses

- Finding the Month in which buyers bought more number of houses.

- Finding relation between Year_built and Price.

Data Pre-Processing

- Dropping the Attributes

- Outlier Analysis

FLOORING AND CAPPING¶

- Data Standardization

- Segregation of Predictors and Target Attributes.

- Splitting the Data into Training set and Test Set

Modeling

- Model Training and Getting Accuracies

- Model Selection

- Feature Selection

1. Checking Feature Importance

2. Checking the accuracies of the model based on different threshold values for the top importance scores.

3. Setting the Threshold value and Selecting the Features.

- Model Tuning and Feature Reduction

- Checking the Performance of Final Model

Pickling the selected model.

Predicting New Data

Conclusion